Artificial Intelligence and Expert Systems

Artificial Intelligence and Expert Systems

Leroy Garcia

1

Definition of AI

Artificial Intelligence is the branch of computer science that is concerned with the automation of intelligent behavior (Luger, 2008).

2

Different Approaches to AI

Systems that think like humans

Systems that think rational

Systems that act like humans Systems that act rational

3

History

Aristotle Rene Descartes Frances Bacon John Locke David Hume Ludwig Wittgenstein Bertrand Russell Rudolf Carnap Carl Hempel Alan Turing

4

Alan Turing

Wrote “Computer Machinery and Intelligence”.

The Turing Test

5

Dartmouth 1956

Automatic ComputersHow can computers be programmed

to use a language?Neuron NetsTheory of the Size of a CalculationSelf-Improvement (Machine

Learning)AbstractionsRandomness and Creativity

6

Acting Human

Natural Language ProcessingKnowledge RepresentationAutomated ReasoningMachine Learning

7

Rational Agents

Anything that can be viewed as perceiving it’s environment through sensors and acting upon it’s environment through actuators.

(Russell & Norvig, 2003)

8

Rational Agents cont.

9

PEAS

Performance Measure EnvironmentActuatorsSensors

Task Environment Made up of PEAS.

10

Examples of Agent Types and PEAS

11

Dimensions of a Task EnvironmentFully Observable vs. Partially

ObservableDeterministic vs. StochasticEpisodic vs. SequentialStatic vs. DynamicDiscrete vs. ContinuousSingle Agent vs. Multiagent

12

Fully Observable vs. Partially Observable

Fully Observable Sensors must provide a complete state

of environment.

Partially Observable Usually due to poor an inaccurate

sensors or if parts of the world are missing the sensor’s data.

13

Deterministic vs. StochasticDeterministic

The action of the next state depends on the action of the previous state.

Stochastic Actions do not depend on previous state.

14

Episodic vs. Sequential

Episodic Single actions are performed.

Sequential Future decisions are determined by the

current action.

15

Static vs. Dynamic

Static Does not change during an agent’s

deliberation.

Dynamic Able to change during an agent’s

deliberation.

16

Discrete vs. Continuous

Discrete Contains finite number of distinct states

and a discrete state of percepts and actions.

Continuous Contains a range of continuous values

17

Single Agent vs. MultiagentSingle Agent

One agent is needed to execute an action on a given environment.

Multiagent More than one agent is needed to

execute an action on a given environment.

18

Examples of Agents and Task Environments

19

Types of Agents

Simple Reflex AgentModel Based Reflex AgentGoal Based AgentUtility AgentLearning AgentProblem Solving Agent

20

Simple Reflex Agent

Selects action based on the current percept and pays no attention to any previous percept.

21

Model Based Reflex Agent Maintains at least some form of internal state

that depends on the percept history and thereby reflects some of the unobserved aspects of the current state.

22

Goal Based Agent

Performs actions based on a specific goal.

23

Utility Based Agent

Takes into account it’s current environment and decides to act on an action that simply makes it happier.

24

Learning Agent

Learning ElementPerformance ElementCriticProblem Generator

25

Learning Agent cont.

26

Problem Solving Agent

State Space Initial StateSuccessor FunctionGoal TestPath Cost

27

Example of States

28

Searching for Solutions

Search Tree States Parent Node Action Path Cost Depth

29

Evaluation of a Search

CompletenessOptimalityTime ComplexitySpace Complexity

30

Various Types of SearchesBreadth-First SearchUniform-Cost SearchDepth-First SearchDepth-Limited Search Iterative Deepening Depth-First

Search

31

Breadth-First Search

Expands the root node first, then all the root node successors are expanded followed by other successors.

32

Uniform-Cost Search

Expands a node with the lowest path cost.

Only cares about the total cost and does not care about the number of steps a path has.

33

Depth-First Search Expands the deepest node and the current fringe

of the search tree. Implements a last-in-first-out methodology.

34

Depth-Limited Search

Solves infinite path problems and can be implemented as a single modification to the general tree search algorithm by setting a depth limit.

35

Iterative Deepening Depth-First Search

Is used to find the best Depth Limit.A goal is found when a Depth Limit

reaches the depth of the shallowest node.

36

Iterative Deepening Depth-First Search cont.

37

Questions?

Any Questions on AI?

38

Expert Systems

Definition “An expert system is an interactive

computer-based decision tool that uses both facts and heuristics to solve difficult decision problems based on the knowledge acquired from an expert.”(The Fundamentals of Expert Systems)

39

Types of Expert Systems

Interpreting and IdentifyingPredictingDiagnosingDesigningPlanningMonitoringDebugging and Testing Instructing and TrainingControlling

40

Creating an Expert System

41

Expert System Programming Languages

PROLOGLISP

42

Mandatory Characteristics

Efficient mix of integer and real variables Good memory-management procedures Extensive data-manipulation routines Incremental compilation Tagged memory architecture Optimization of the systems

environment Efficient search procedures

43

Levels of an Expert System Knowledge base

Problem-solving rules, procedures, and intrinsic data relevant to the problem domain.

Working memory Task-specific data for the problem under

consideration.   Inference engine

Generic control mechanism that applies the axiomatic knowledge in the knowledge base to the task-specific data to arrive at some solution or conclusion.

44

Expert Systems Organizational and Operating Environment

45

Data Flow of an Expert System

46

References  Expert Systems: Wikipedia. (n.d.). Retrieved October 18, 2008, from

Wikipedia: wikipedia - http://en.wikipedia.org/wiki/Expert_system  Fogel, D. B. (2002). Blondie24: Playing at the Edge of AI. San

Fransisco,CA: Morgan Kaufman Publishers.  Luger, G. F. (2008). Artificial Intelligence. Boston: Pearson Addison

Wesley.  Russell, S., & Norvig, P. (2003). Artificial Intelligence: A Modern

Approach. Upper Saddle River, NJ: Pearson Education Inc.

The Fundamentals of Expert Systems. (n.d.). Retrieved November 13, 2008, from http://media.wiley.com/product_data/excerpt/18/04712933/0471293318.pdf

47

The End

Any Questions?

48